MicroRNAs (miRNAs) are small double-stranded RNAs that, in vertebrates, function as negative regulators of gene expression by base pairing with the 3’-UTR of target mRNAs while part of RNA-induced silencing complexes (RISCs). More than 400 miRNAs have been identified in both mice and humans, each of them having the potential to regulate hundreds of target genes. It has been proposed that more than one third of all human genes may be regulated by miRNAs. Most of experimentally detectable miRNAs are expressed in the brain and, among those upregulated in a tissue specific expression pattern in the embryo, half are brain-specific/enriched.
miRNAs are processed by Dicer, a dsRNA-specific endonuclease, from an inactive ~70-nt precursor RNA to a functional ~25-nt molecule.
The development of therapies for inherited retinal degenerations relies upon a better understanding of the cell biology of this heterogeneous family of diseases. In particular, it is of relevance to study alterations occurring in the retina of various lines of transgenic, knock-out and spontaneous-mutant mice, which mimic various types of retinal degeneration.
For this reason, the first aim of this project is to characterize the retinal phenotype in a mouse in which the production of retinal microRNAs has been greatly impaired by knocking out Dicer, the gene encoding for the key enzyme for the synthesis of miRNAs. Our final goal is to test whether miRNAs could represent a novel class of candidate disease genes.
Dicer null mice have been reported to die at embryonic day 7.5 rendering it impossible to study the role Dicer may play in later stages of development or in adult tissues. To bypass that problem, a floxed Dicer allele has been recently created. Using this allele in conjunction with a retinal-specific cre (Chx10-Cre) allele, it has been possible to inactivate Dicer in the retina selectively, thus creating a conditional knock-out (CKO) mouse. Since Dicer is required for the processing of miRNAs, removal of this enzyme results in a significant decrease in mature miRNAs.
Retinal sections from CKO and wt mice aged postnatal (P) 16, 30, 45 days and 3, 4, 5, and 7 months were processed for immunocytochemistry to cell type specific antibodies. Eyes used for morphological analysis were from animals previously used for ERG recordings. Moreover, retinas from CKO and wt animals of different ages were processed by Northern blot analysis to confirm the decrease of mature microRNA levels.
As previously reported, the Chx10Cre transgenic allele does not drive Cre expression in all retinal cells; thus, not all cells in CKO retinas are expected to lack a functional Dicer protein. For this, we performed in situ hybridization on retinal sections with a probe specific for the floxed exon in the Dicer conditional allele. We found that native Dicer is expressed in the vast majority of retinal cells of control mice, while in the CKO retinas Dicer expression is patchy, with groups of labeled cells interdigitated with unlabeled cells.
Northern blot and in situ hybridization analysis confirmed that Dicer CKO retinas are indeed deficient in mature miRNAs (miR-183, miR-124a and miR-96).
Development of Dicer CKO retinas is grossly normal. All cell types analyzed are present at postnatal day 16 (P16). However, all retinas examined show the presence of rosettes, typical circular structures in which photoreceptors are oriented toward an internal lumen. Rosettes have previously been described in retinoblastoma, diabetic retinopathy and retinitis pigmentosa in association with retinal degeneration and/or abnormal proliferation. Nevertheless, BrdU and phosphohistone H3 staining fails to detect dividing cells in Dicer CKO retinas. Thus, rosette formation is not caused by increased proliferation of retinal cells.
At P16, rosettes are scattered along an otherwise normal retinal surface and are primarily composed of photoreceptors and their synaptic terminals.
The time period from P16 to P45 is characterized by progressive alteration and remodelling of the laminar retinal structure: the number of rosettes increases and their structure becomes more complex, with migration of second order neurons into the outer nuclear layer.
At 3 months, rosettes decrease in size and eventually disappear. Photoreceptors composing the rosettes degenerate and the outer nuclear layer becomes progressively thinner. DNA-binding dyes reveal pycnotic nuclei in the outer and inner retina. Simultaneously, Müller cells become hypertrophic: radial processes increase in size and GFAP reactivity raises, indicating a generalized glial activation. At the same time, Cre positive cells become scarce, possibly because a large fraction of these Dicer null cells degenerate. Surviving rod bipolars are often in clusters and have altered morphologies with hypertrophic axonal arborizations. Most of the residual rod bipolar cells are Cre-negative and therefore most likely originated from precursors in which Dicer was not inactivated.
The retinal-specific Dicer-KO mouse we studied is the first example of miRNAs inactivation in an area of the mammalian Central Nervous System and demonstrates that such a genetic manipulation leads to an aberrant retinal layering eventually culminating with extensive retinal degeneration.
Summarizing, inactivation of Dicer results in a decrease of retinal-expressed miRNAs. Dicer CKO animals show a unique pattern of progressive retinal disorganization and degeneration. These results represent the first step towards identifying the role miRNAs contribute to retinal cell function and survival. The reported data also provide insight into the molecular mechanisms of retinal degeneration and the maintenance of normal retinal function and architecture and point out to miRNAs as likely candidate disease genes.
Retinal degeneration in Dicer CKO mice does not parallel precisely other known retinal pathologies. However, some similarities could be noticed. For instance, as occurs in retinitis pigmentosa (RP), photoreceptors are the first neurons to be hit by miRNA withdrawal. In addition, in both RP and Dicer CKO mice inner retinal neurons like ganglion cells appear to be preserved.
Retinitis Pigmentosa is a family of inherited diseases leading to photoreceptor death and represents one of the major causes of blindness in the world.
Among therapeutic strategies under study, the devise and implant of electronic prostheses has reached the stage of clinical trials. Retinal ganglion cells (RGCs) form the biological substrate that epiretinal prostheses stimulate to restore vision in RP; hence, the efficacy of such electronic implants depends upon RGCs viability. For all these reasons, we focused our attention on
the morphology and survival of RGCs in the rd1 mouse, an established model of autosomic recessive RP, at various stages from the onset of photoreceptor death.
In order to study single RGCs, we generated a transgenic rd1/Thy1-GFP-M mouse line, crossing GFP-M mice, expressing GFP in a small number of heterogeneous RGCs of various types, with rd1 mice, in which rod death starts in the 2nd week of life. As the two founder strains are on C57BL/6J and C3H backgrounds, respectively, repeated backcrossing was performed to isolate double mutants free of background effects. We are presently studying the F9 generation. Retinal whole mounts from animals aged 1 to 12 months were stained with anti GFP antibodies and examined by confocal microscopy. After classification, each cell was studied quantitatively after Neurolucida tracings and Metamorph analysis. To estimate survival in the ganglion cell layer (GCL), retinal whole mounts were stained with ethidium and cell nuclei counted in serial optical sections.
The GCL in rd1 mutant mice aged 1 year displayed at a first sight clear signs of reorganization. Cell bodies were irregularly spaced, with empty areas suggestive of neuronal degeneration. Often cells appeared arranged in circular domains resembling the initial stage of rosette formation. In addition, chromatin condensation was observed in a considerable number of cells.
Extensive cell counting in rd1 mice aged 1 year reveals excellent survival of cells in the ganglion cell layer (94.6%). Nevertheless, a dramatic increase in the number of condensed nuclei (+ 85%) was observed in these retinas with respect to a previous time point at 4.5 months.
In the novel rd1/Thy-1 GFP strain, single RGCs exhibiting GFP labeling occurred in lower number as compared to the wild type founder. This characteristic is not caused by degeneration of the ganglion cell layer, but simply for the mice analyzed were heterozigous unlike the founder. We empirically discovered that the number of green ganglion cells is sensitive to the number of alleles of the Thy-1 GFP transgene.
Our studies suggest that these neurons have also undergone cell body shrinkage and dendritic tree retraction. This finding could be misleading, as it’s very similar to regressive remodeling previously described for bipolar cells as an effect of deafferentation. By measuring ganglion cells from 1 month-old mice, we discovered how the problem described is attributable more to an unrealized postnatal growth rather than a remodeling for retinal degeneration.
This study is telling us how the postnatal development of their dendritic trees is affected in the rd1/Thy-1 GFP mouse, a new strain of mouse dedicated to the study of retinal ganglion cells in retinal degeneration.
The results of these two projects together give light to different mechanism of plasticity and remodelling of retinal cells in postnatal development and adult life of normal and pathological mammalian models.